The DeWaTra COST Action (CA23159) – Decarbonising Waterborne Transportation is pleased to announce the Operational Vessel Data Challenge, presented and sponsored by DNV in collaboration with vessel data partner Miros.
The challenge invites researchers and industry experts to develop and benchmark vessel-performance modelling approaches using a common set of real operational and environmental data.
Turning operational vessel data into trustworthy efficiency insights
Energy efficiency is expected to remain one of the maritime sector’s most important near-term tools for reducing fuel consumption and emissions. However, verifying efficiency improvements under real operating conditions remains challenging.
Different vessels and fleets may use different sensors, preprocessing methods and modelling assumptions, making performance results difficult to compare. The Operational Vessel Data Challenge therefore aims to create a common environment in which different modelling approaches can be assessed consistently.
Participants will develop models designed to reproduce vessel speed–power behaviour using operational and metocean data. Participants may also explore the ability of their models to predict future behaviour and generalise to new datasets.
Approaches may include physics-guided models, data-driven methods, artificial intelligence and machine learning, or hybrid approaches.
The dataset
Registered teams will receive access to an anonymised dataset containing several months of synchronised vessel and environmental measurements.
The available parameters include, among others:
• Shaft speed and shaft power
• Speed Through Water and Speed Over Ground
• Fore and aft draft and displacement information
• Vessel heading
• Significant wave height, wave period and direction
• Wave directional spectrum
• True and relative wind speed and direction
• Current speed and direction
• Sea surface temperature
• Salinity
• Water depth
From September 2026, registered teams will receive the link to access the challenge data through the DNV Veracity platform.
What will be assessed?
Submissions will be evaluated according to three main areas:
Quality of modelling, including the ability to capture transient behaviour, sensitivity analysis, uncertainty estimation and prediction for new datasets;
Originality of the approach, including adaptation or improvement of existing models and innovative combinations of physics and AI approaches; and
Clarity of the submitted solution, supported by a common reporting template.
Through the challenge, the organisers also aim to identify which inputs are genuinely necessary, how strongly preprocessing influences results, where models fail and how uncertainty should be addressed.
These findings could contribute to more standardised measurement–model workflows and help make vessel-performance results more comparable and trustworthy for future operational decision support.
Key dates
• Challenge opening: 25 May 2026
• Data access: From September 2026
• First online Q&A session: October 2026
• Second online Q&A session: February 2027
• Submission deadline: 31 August 2027
• Winner announcement: Q4 2027
• Winner visit to DNV headquarters: 2028
Award and collaboration opportunity
The winning team will be invited to spend one week at DNV headquarters in Norway, with opportunities to meet representatives from DNV Maritime Research, Maritime Business Development & Innovation, Maritime Advisory and experts in maritime sensing technologies.
Travel, hotel accommodation and meals will be provided.
Following the challenge, the submitted approaches will also provide valuable input for a review of the methodologies, their performance, strengths and limitations.
How to participate:
Interested participants should register through the DeWaTra challenge webpage. One registration email is required per team.
From September 2026, registered teams will receive information and the link required to access the challenge dataset through DNV Veracity.
👉 Register and learn more: https://dewatra.eu/operational-vessel-data-challenge/
This Action is supported by COST – European Cooperation in Science and Technology, Grant CA23159.